Agreement between cardiovascular disease risk assessment tools in the Arabian gulf countries population

A Oulhaj, A Shehab, K Al Rasadi, W Mahmeed

Research output: Contribution to journalArticle

Abstract

Background: Cardiovascular disease (CVD) is the leading cause of morbidity and mortality in the Arabian Gulf countries (AGC), accounting for approximately 45% of all deaths in these predominantly young population. To date, no CVD risk assessment tool has been specifically developed or properly calibrated to be used for the AGC population. Instead, a variety of existing CVD risk prediction tools derived from regions other than AGC are used with no evidence on which prediction tool performs well. Purpose(s): We aim to investigate, based on a large data set from the AGC population, the agreement between predicted CVD risks derived from a variety of externally validated risk prediction tools. Method(s): The agreement analysis of predicted risks was carried out using a data set from the Centralized Pan-Middle East Survey (CEPHEUS), a large cross-sectional study involving 5274 patients in six Arabian Gulf countries (Bahrain, Oman, Qatar, the United Arab Emirates, the Kingdom of Saudi Arabia, and Kuwait). The list of risk prediction tools includes the systematic coronary risk evaluation (SCORE), The ACC/AHA Pooled Cohort risk equations (PCRE), the National Cholesterol Education Program Framingham score (NCEP-ATP-III) and the laboratory Framingham risk (FRAM). We calculated the overall and pairwise Lin's concordance correlation coefficient (CCC) to measure the degree of agreement between all, and each pair of, risk tools. We also calculated, for each risk tool, the percentage of subjects assigned to the high risk category according to the thresholds provided in their original guidelines. Result(s): To enable a fair comparison across different risk prediction tools, only subjects without previous history of CVD and those aged 45 to 65 years were used in the final analysis (n = 2271). Overall, we found a poor agreement between the different risk prediction tools (overall CCC= 0.32). When compared to SCORE, the Lin's CCC was 0.40, 0.26, 0.11 and 0.11, for respectively NCEP-ATP-III, PCRE (white ethnicity), PCRE (African-American ethnicity) and FRAM. Furthermore, PCRE for African-American ethnicity produced the highest proportion of participants at high risk (72%). This proportion according to the remaining tools was 42% for PCRE (White ethnicity), 28% for FRAM, 9% for SCORE and 3% for NCEP-ATP-III. Among the 2060 participants who were classified as non-high risk patients according to SCORE, 69%, 36%, 22% and 1% of them were re-classified as high risk according to PCRE (African-American ethnicity), PCRE (White ethnicity), FRAM and NCEP-ATP-III. Conclusion(s): We showed a poor agreement, in the AGC population, between a variety of externally validated CVD risk assessment tools. This demonstrates the difficulty of choosing any of these tools for public health and clinical interventions in this region. Due to the high burden of CVD in the AGC population, there is an urgent need to improve the evidence base of CVD risk assessment tools in this region.
Original languageEnglish
Pages (from-to)S60-S61
JournalEuropean Journal of Preventive Cardiology
Volume25
Issue number2
Publication statusPublished - 2018

Keywords

  • *cardiovascular disease
  • *risk assessment
  • African American
  • Bahrain
  • Cardiovascular Diseases
  • Framingham risk score
  • Kuwait
  • Oman
  • Qatar
  • Risk Assessment
  • Saudi Arabia
  • United Arab Emirates
  • adult
  • aged
  • cholesterol
  • cohort analysis
  • conference abstract
  • controlled study
  • coronary risk
  • correlation coefficient
  • cross-sectional study
  • education program
  • ethnicity
  • female
  • human
  • major clinical study
  • male
  • practice guideline
  • prediction
  • public health

Fingerprint Dive into the research topics of 'Agreement between cardiovascular disease risk assessment tools in the Arabian gulf countries population'. Together they form a unique fingerprint.

Cite this